For the most part, viral vector–based gene therapies have been used against rare diseases. But now these gene therapies are also starting to be used against diseases that have larger patient populations. As a result, there is rising demand for viral vectors, mainly for adeno-associated virus (AAV) vectors. It is pressuring the AAV manufacturers—as well as the companies that supply the AAV manufacturers with production and purification systems—to address a range of practical problems.

A key problem in manufacturing is the lack of standardization. Multiple upstream and downstream platforms and processes are emerging as developers progress from research-scale to commercial-scale production. Proliferating platforms and processes can lead to additional problems, such as regulatory complications. As welcome as consolidation might be, it seems a distant prospect. Different stakeholders have different needs.

One of the first decisions a therapeutic developer needs to make is whether to develop its own AAV manufacturing infrastructure (a costly and time-consuming endeavor) or to outsource. Companies looking to outsource development and manufacturing activities for AAV vector–based therapeutics need to assess their internal capabilities and to determine how these capabilities may be complemented by the expertise of partners, such as contract development and manufacturing organizations.

Customized solutions

“We offer customized AAV development and manufacturing solutions to meet our customers’ timelines and minimize risks,” says Michael Baker, senior director, Viral Gene Therapy, Fujifilm Diosynth Biotechnologies. “Our goal is to deliver a complete CMC [chemistry, manufacturing, and control] package to enable investigational new drug filings and subsequently allow the product to enter the clinic. For example, our triple transfection system utilizes a clonally selected cell line and optimized plasmids to produce industry-leading titers and packing efficiencies.”

Customer genes of interest can be used interchangeably with the company’s plasmid constructs. Both the cell line and plasmids are off-the-shelf materials, removing the lengthy plasmid procurement process, and they are offered as both research and good manufacturing practice (GMP) grades on a pay-as-you-use basis.

“We have optimized our process to work with all major serotypes with minor tweaks, enabling a streamlined development approach with minimal need for upfront work,” Baker asserts. “This reduces the length of process development and enables customers to go from gene sequence to fully released GMP material in 16 months, which includes DP [drug product], DP release, as well as an engineering run.”

Both on- and off-platform solutions are offered with demonstrated experience up to 2 × 500 L. “We provide full end-to-end solutions,” Baker continues, “and we offer known scalability with regard to process-related infrastructure, independent of the platform.”

Baker recalls how Fujifilm assisted a client that had a very high dose requirement. This client also had challenges with the productivity of other triple transfection systems and residual DNA in the final product. Fujifilm utilized its clonal cell line and optimized plasmids, cloned the customer’s gene of interest construct into a plasmid backbone to generate the AAV vector, and achieved 2–5 × 1011 vg/mL titers at harvest.

Two different optimizations were used to treat the resultant vectors to maximize DNA removal. Both methods showed a 2-log reduction of host cell DNA when compared to standard conditions and enabled progression of the client’s clinical program.

Upstream improvements, downstream benefits

Oxford Biomedica Solutions suggests that its “plug and play” platform for AAV purification can expedite the development of new constructs (or capsids) and the delivery of high-quality vectors. “Our platform was created by developing a deep understanding of chromatographic behavior, scalability, and the impact of upstream construct design and bioreactor performance on downstream purification, all with a focus on demonstrating process robustness across scales and products,” says Ashish Sharma, a senior scientist at the company.

“Optimized vector design with dual transfection and optimized bioreactor operation increased packaging efficiency and bioreactor productivity,” Sharma continues. “We achieved > 1 E + 15 vg/L titers. This high-titer achievement then enabled the intensification of the purification process.”

The high-titer bioreactor productivity allowed for direct capture on affinity chromatography, with removal of a tangential flow filtration volume reduction step. This chromatography operation can now be accomplished in one manufacturing shift. These two improvements resulted in significant time and cost savings.

Construct design and bioreactor packaging efficiency has a large impact on the anion-exchange (AEX) step. Proper construct design allows for better packaging efficiency, which then reduces the empty capsid burden on the AEX step. “This lower empty capsid load combined with extensive AEX design space development through a DOE [design of experiment] approach led the platform to deliver scalable and consistent product quality,” Sharma asserts.

Development robustness is measured through process and product quality conformance to historical performance ranges and trends for every unit operation across scales, a typical benchmark.

“An example of our platform’s robustness is that we have successfully executed >45 batches in our GMP facility,” Sharma points out. “Our general approach can be applied toward multiple AAV serotypes.

“We have consistently achieved 1 E + 15 vg/mL titers and <10% empty capsids across multiple serotypes. The high titers with our efficient purification process is able to deliver >1 E + 17 vg/500 L batch. We are confident that the use of our platform is an important step to further democratize AAV gene therapy, and that it will provide our partners with much better cost-of-goods economics.”

Removal of empty capsids

Upstream optimization is imperative for generation of a controlled empty/full capsid ratio. Although empty capsids have the potential to increase antigen load, they may also serve as decoys. That is, they may bind antibodies and make it easier for full capsids to deliver therapeutic genes to target cells. An ideal vector preparation may contain decoy capsids (provided they have been modified to be less visible to the immune system) as well as blockers of innate immune sensors or stimulators of T regulatory cells. (Hoffmann and Herzog. Mol Ther. 2013; 21(9): 1648–1650).

According to Aydin Kavara, PhD, team leader, R&D Bioprocessing, Pall Corporation, the challenges posed by empty capsids are particularly striking when difficult-to-scale methods are used for the development of purification processes and the generation of initial batches of viral vector material. When chromatography methods are applied using a linear conductivity gradient, incomplete peak resolution can complicate “peak cutting.” Simple mathematical scale-up models can be untenable.

Empty capsids are the same size as full capsids. However, empty capsids and full capsids typically have different densities and isoelectric points, potentially allowing separations to be conducted on the basis of ultracentrifugation and chromatography, respectively. The problem, however, is that density differences and isoelectric point differences may be too small to allow empty and full capsids to be easily separated. “The problem becomes even more complex when we factor in the presence of ‘partial’ or ‘overpacked’ capsids,” Kavara points out. “These capsids negatively impact product efficacy and safety.”

On scale-up, upstream batch-to-batch reproducibility problems are amplified as material is passed downstream. “This necessitates the availability of robust downstream methods that do not operate on the edge of failure and can handle upstream variability,” Kavara remarks.

Such a method, he suggests, is possible through the use of Pall’s Mustang Q XT chromatography capsules. Instead of traditional linear gradients, the capsules employ small (about 1 mS/cm) conductivity steps. According to Pall, the small steps result in a series of discrete elution peaks, simplifying assessment of the purification and thereby accelerating process development.

The empty capsids predominantly elute in the first peak and the full capsids in the second peak. “When the initial ratio of full to empty capsid and other impurities is large, approximately over 20%, we can achieve an elution of almost pure full capsids,” Kavara asserts.

A similar approach is possible with Capto Q resin. Nonetheless, Kavara believes that the Mustang Q XT capsules offer certain advantages, including “the convective nature of flow on membranes, plug-and-play availability, and ease of use.” With these advantages, the capsules can offer greater “speed, resolution, and cost effectiveness.”

Scalability issues

According to a report from McKinsey & Company, AAV vectors are used in 82% of the viral vector–based gene therapies in the pipeline. To address high (and growing) demand for AAV vectors, Thermo Fisher Scientific developed the Gibco AAV-MAX Helper-Free AAV Production System. It is designed to enable gene therapy developers to scale production and speed gene therapy candidates through product pipelines. By delivering more viral particles per production volume, the all-in-one system helps to increase efficiency and reduce AAV production cost.

Gibco AAV-MAX Helper-Free AAV Production System chart
Thermo Fisher Scientific developed the Gibco AAV-MAX Helper-Free AAV Production System to enable gene therapy developers to scale production. The AAV-MAX system yields high titers across multiple AAV serotypes. The system was used to produce six AAV serotypes at 30-mL scale in 125-mL shake flasks. Titers were measured by qPCR.

“The AAV-MAX System is the only complete, fit-for-purpose solution available that is manufactured under cGMP conditions to enable large-scale AAV production,” says Emily Jackson-Holmes, PhD, product manager, Thermo Fisher Scientific. “It includes traceability and regulatory documents, helping customers seamlessly transition from research to large-scale commercial production.”

Several challenges are associated with existing AAV production systems, including low titers, the high cost of cGMP plasmid DNA, and a lack of fit-for-purpose reagents. “We developed the AAV-MAX System to address these challenges by using preoptimized and regulatory-compliant reagents fit for commercial production to deliver higher AAV yield,” Jackson-Holmes notes. “This can mean significant savings—up to 50% on upstream AAV production costs and 25% on plasmid DNA costs.”

“We continue to hear that scalability is one of the biggest hurdles, so we developed a system that could grow as clinical pipelines mature,” Jackson-Holmes adds. “With the RUO [research use only] and CTS [cell therapy system] versions of the AAV-MAX system, we can help streamline the transition from research applications to large-scale clinical manufacturing.

“There is a desire to switch from adherent-based to suspension-based AAV production systems that use regulatory-friendly reagents and address challenges with scalability. Customers also want to improve existing AAV platforms to achieve higher productivity. They have obtained better yields with the AAV-MAX System at scales up to 200 L-plus compared with their current platforms and solutions.”

Speed to market

MilliporeSigma provides viral vector manufacturing platforms that are designed to simplify upstream workflows and enable processes that are easier to manage, adjust, and scale. For example, last August, the company extended its VirusExpress offering by introducing the VirusExpress 293 Adeno-Associated Virus (AAV) Production Platform.

“The VirusExpress 293 AAV Production Platform enables biopharmaceutical companies to increase the speed to clinical manufacturing while reducing process development time as well as costs,” states Dirk T. Lange, executive vice president, global head, Life Science Services, MilliporeSigma.

“With this platform’s launch, we are one of the first CTDMOs [contract testing, development, and manufacturing organizations] and technology developers to provide a full viral vector manufacturing offering,” Lange continues. “With more than 30 years of drug development, material sciences, and process technology expertise, our Millipore CTDMO Services span preclinical development to commercial manufacturing, including testing, across the modalities with the most patient impact.”

In addition to accelerating process development, the suspension culture format allows for larger batch yields. Suspension culture is amenable to scalable robust manufacturing processes, while being less labor intensive. The chemically defined cell culture media eliminates the safety, regulatory, and supply chain concerns related to animal- and human-derived materials.

Lange notes that another VirusExpress offering, the VirusExpress Lentiviral Production Platform, has allowed customers to reach clinical manufacturing in as little as six months, to achieve titers that are five times higher than those of the leading competitor, and to transition legacy processes to templated solutions.